[tag: Load and preprocess images]
  images = []
  for image_path in image_paths:
      img = cv2.imread(image_path, cv2.IMREAD_GRAYSCALE)
      img = cv2.resize(img, (128, 32))
      images.append(img)
 
  images = np.array(images) / 255.0
  images = np.expand_dims(images, axis=-1)  ** # Add channel dimension for grayscale images**

  label_set = list(set(labels))  # Convert labels to a set and then a list
  label_to_index = {label: index for index, label in enumerate(label_set)}
  labels = [label_to_index[label] for label in labels]

 [tag: Convert labels to one-hot encoding]
 max_label_len = max([len(str(label)) for label in labels])  # Adjust this as per your needs
 labels_padded = np.zeros((len(labels), max_label_len))
 for i, label in enumerate(labels):
      labels_padded[i, :len(str(label))] = list(str(label))
 [tag: input format]
 input_img = Input(shape=(32, 128, 1), name='image_input')
 labels_input = Input(shape=(max_label_len,), name='label_input')
 input_length = Input(shape=(1,), name='input_length')
 label_length = Input(shape=(1,), name='label_length')
 
 

above code is the input for CNN layers ,I need improvement in that section

 [tag:CNN layers]
 x = Conv2D(32, (3, 3), activation='relu', padding='same')(input_img)
 x = MaxPooling2D((2, 2))(x)
 x = Conv2D(64, (3, 3), activation='relu', padding='same')(x)
 x = MaxPooling2D((2, 2))(x)

 ```This code giving me error at the time of training the dataset so I need to make some    changes in the above code ``` 
new_shape = (8, 2048)  # Ensure the reshaped dimensions match the total size of the tensor
x = Reshape(target_shape=new_shape)(x)  # Adjust shape to match the output tensor size
x = Bidirectional(LSTM(128, return_sequences=True))(x)
x = Bidirectional(LSTM(128, return_sequences=True))(x)
y_pred = Dense(len(label_set) + 1, activation='softmax', name='y_pred')(x)  # +1 for CTC blank       label
#Dense layer with softmax activation

def ctc_lambda_func(args):
y_true, y_pred, input_length, label_length = args
return tf.keras.backend.ctc_batch_cost(y_true, y_pred, input_length, label_length)

After Using above code getting following error

`—————————————————————————
ValueError Traceback (most recent call last)
Cell In[23], line 5
2 train_outputs = {‘ctc’: np.zeros((len(images),))}
4 # Train the model
—-> 5 ctc_model.fit(train_inputs, train_outputs, epochs=10, batch_size=32)
6 # Dummy data for CTC
7 train_outputs = {‘ctc’: np.zeros((len(images),))}

File ~AppDataLocalProgramsPythonPython312Libsite-packageskerassrcutilstraceback_utils.py:122, in filter_traceback..error_handler(*args, **kwargs)
119 filtered_tb = _process_traceback_frames(e.traceback)
120 # To get the full stack trace, call:
121 # keras.config.disable_traceback_filtering()
–> 122 raise e.with_traceback(filtered_tb) from None
123 finally:
124 del filtered_tb

Cell In[18], line 4, in ctc_lambda_func(args)
2 def ctc_lambda_func(args):
3 y_true, y_pred, input_length, label_length = args
—-> 4 return tf.keras.backend.ctc_batch_cost(y_true, y_pred, input_length, label_length)

ValueError: Exception encountered when calling Lambda.call().

Input tensor functional_1_1/ctc_1/Cast_2:0 enters the loop with shape (1, 1), but has shape (1, None) after one iteration. To allow the shape to vary across iterations, use the shape_invariants argument of tf.while_loop to specify a less-specific shape.

Arguments received by Lambda.call():
• inputs=[‘tf.Tensor(shape=(None, 1), dtype=float32)’, ‘tf.Tensor(shape=(None, 8, 4), dtype=float32)’, ‘tf.Tensor(shape=(None, 1), dtype=float32)’, ‘tf.Tensor(shape=(None, 1), dtype=float32)’]
• mask=[‘None’, ‘None’, ‘None’, ‘None’]
• training=True`